Teach Computer To See
In the field of video-based emotion recognition, analyses of facial expressions are most common. Here, emotion is often expressed by subtle changes in one or a small set of discrete facial features. For example, anger might be displayed by a tightening of the lips or sadness by lowering the corners of the mouth. In general, facial expressions are based on a change in various muscular action units. The action units can be analyzed individually as well as in combination. The changes can mark a range of emotions, such as sadness, depression, fear, and anger.
With the increase in automation much progress has been made in the development of computer systems which understand this human expression of. Moreover, analyses of facial muscle activity were not only found to indicate different emotions, but also psychological disorders. For instance, persons with high depression symptom severity were found to display contempt, less smiles, and smiles related to contempt. However, head movements and body gestures can be taken into account as well adding further markers for automatic emotion recognition.